High Contrast and Detail Effect - 16 or 8 Bit?

[Boatman]

I've taken quite a liking to the "High Contrast and Detail Effect" method of increasing photo sharpness and contrast explained by forum member Jiro in the article, High Contrast and Detail Effect using dodge and burn technique.. I think I apply it to probably 75% of my images now, at least in some low opacity level.

I normally try to work in 16-bit mode though I confess I often fail to turn this on since ACR opens into Photoshop in the 8-bit mode automatically. (Does anyone know how to change this?) When using Jiro's technique in 16-bit mode, the application of the blur layer takes about ten minutes - seriously! My computer is not really up to date, but it does have an Intel Q6600, 4-core, 2.4GHz CPU, so it is no slouch.

In 8-bit mode, applying the blur layer takes about 30 seconds. So if I'm using 16-bit, usually before applying the blur layer I'll switch to 8-bit and then switch back to 16-bit before doing additional adjustments. What impact does changing the bit depth have on the actual blur process? Is this a good or bad thing to do from an image quality standpoint? Does switching back and forth between 8 and 16-bit within an image create any issues?

[Colin Southern]

I normally try to work in 16-bit mode though I confess I often fail to turn this on since ACR opens into Photoshop in the 8-bit mode automatically. (Does anyone know how to change this?)

Yes

There's a hyper-link looking thing at the bottom of the ACR screen - just click that and you can set things like 8/16 bits / colourspace / whether or not sharpening is applied etc.

[Kris V]

I was under the impression that once a picture has been set to 8 bit mode, changing to 16 bit only increases file size, but does not improve the picture quality any more.
True or false?
Once I have converted to 8 bit (some filters I have won't work in 16 bit images), I leave them like that. Does it make sense to convert back to 16 bit?

[Colin Southern]

I was under the impression that once a picture has been set to 8 bit mode, changing to 16 bit only increases file size, but does not improve the picture quality any more.
True or false?
Once I have converted to 8 bit (some filters I have won't work in 16 bit images), I leave them like that. Does it make sense to convert back to 16 bit?

Hi Kris,

The biggest advantage of 16 bit is when you're pushing and pulling tonal ranges and you want to avoid banding and other nasties caused by steps that become so big in an 8 bit scheme that they become visible.

So if you get to the point where you've already done the damage then switching to 16 bit isn't going to help - or if you're close to the limit under 8 bit - switch to 16, and make more adjustments then even though it's 16 bit, you may still start to see the damage that was done under 8 bit.

But if you switch to 16 bit after opening the image in 8 bit you may still sidestep things like rounding errors etc.

I think the best advice I can give is always work in 16 bit mode if you can (and then convert back to 8 bit at the end if you wish), but switching to 16 bit may still give you more of a safety margin than you'd have had had you have stuck to 8 bit all the way.

The best plan of all though is to do all of your big adjustments whilst the image is still in RAW, for a couple of reasons:

1. The image is still in linear gamma - so no information has been thrown away yet, and

2. What a lot of people don't realise is that the likes of ACT only ever applies one "adjustment" to an image. eg If you raise the exposure - apply a GND - boost the saturation etc, ACR doesn't apply one adjustment - then another - then another - it rolls them all into 1 big one, which gives the image the "best possible chance in life" when it's released from the incubator (ACR) and out into the big wide world (Photoshop)

[Kris V]

Hi Kris,

The biggest advantage of 16 bit is when you're pushing and pulling tonal ranges and you want to avoid banding and other nasties caused by steps that become so big in an 8 bit scheme that they become visible.

So if you get to the point where you've already done the damage then switching to 16 bit isn't going to help - or if you're close to the limit under 8 bit - switch to 16, and make more adjustments then even though it's 16 bit, you may still start to see the damage that was done under 8 bit.

But if you switch to 16 bit after opening the image in 8 bit you may still sidestep things like rounding errors etc.

I think the best advice I can give is always work in 16 bit mode if you can (and then convert back to 8 bit at the end if you wish), but switching to 16 bit may still give you more of a safety margin than you'd have had had you have stuck to 8 bit all the way.

The best plan of all though is to do all of your big adjustments whilst the image is still in RAW, for a couple of reasons:

1. The image is still in linear gamma - so no information has been thrown away yet, and

2. What a lot of people don't realise is that the likes of ACT only ever applies one "adjustment" to an image. eg If you raise the exposure - apply a GND - boost the saturation etc, ACR doesn't apply one adjustment - then another - then another - it rolls them all into 1 big one, which gives the image the "best possible chance in life" when it's released from the incubator (ACR) and out into the big wide world (Photoshop)

Guess I know just about enough to be dangerous.
This sure clarifies things. When I move from LR to CS5, I try to keep working in 16 bit whenever possible, but like I said - some filters are grayed out and if I want to use them, I need to convert. I do try to leave those for the last possible moment.
I don't convert to 8 bit when I don't need to, and then of course end up with huge tiff files!

[Boatman]

...a couple of good pointers here. Thank you all. BTW, I created a task for the Jiro technique that I can run and it does the whole process automatically. I hadn't used that tool before. The blur step still takes a long time to do in 16-bit mode though.